Comment by a12k

1 year ago

Ollama is sketchy enough that I run it in a VM. Which is odd because it would probably take less effort to just run Llama.cpp directly, but VMs are pretty easy so just went that route.

When I see people bring up the sketchiness most of the time the creator responds with the equivalent of shrugs, which imo increases the sketchiness.

It's fully open source. I mean yes it uses llama.cpp without giving it credit. But why run it in a VM?

  • It severely over-permissions itself on my Mac.

    • Can you please elaborate? How are you running ollama? I just build it from source and have written a shell script to start/stop it. It runs under my local user account (I should probably have its own user) and is of course not exposed outside localhost.

    • Just install it from Brew and run the service in a separate terminal tab.

  • > But why run it in a VM?

    Because you don't execute untrusted code in your machine without containerization/virtualization. Don't you?

    • The question was asking why it’s untrusted code, not why you run untrusted code in a VM.

      There are a lot of open-source tools that we have to trust to get anything done on a daily basis.

    • Every single day. There's just too much good software out there, and life is too short to be so paranoid.

> but VMs are pretty easy so just went that route.

Don’t you need at least 2 GPUs in that case and put kernel level passthrough?

ollama advertising llama.cpp features as their own is very dishonest in my opinion.

  • That’s the curse and blessing of open source I guess? I have billion dollar companies running my oss software without giving me anything - but do I gripe about it in public forums? Yea maybe sometimes but it never helps to improve the situation.